There are 800,000 new strokes in the United States each year. Functional improvement of the upper paretic limb after stroke is mainly determined by improvement of the paretic hand, yet restoration of hand function after stroke often lags behind restoration of more proximal joints, and impairments are often resistant to therapeutic intervention. The rationale for our approach stems from the growing evidence that neuro-rehabilitation after stroke may be enhanced via the application of motor learning strategies within the context of repetitive movement practice. The key therapeutic aspects of these strategies are high repetition, volitional effort, and successful completion of tasks to prevent frustration. While these represent promising therapeutic strategies, they are limited to mildly impaired subjects who already have enough control of finger extension to tolerate high repetitions of grasp/release tasks without succumbing to fatigue and frustration. There is a very large population of stroke patients who don't fall into this category and often must rely on inefficient compensatory strategies. In these patients, unassisted repetitive task practice may not be the optimal retraining strategy. Previously we have developed HandSOME (Hand Spring Operated Movement Enabler), a passive exoskeleton that provides extension assistance during a simple grasp pattern. HandSOME is inexpensive, lightweight and can be donned independently by patients. Hand range-of-motion and function is increased immediately when wearing HandSOME and enables successful practice of reach and grasp tasks, even in subjects with severe hand impairment. The specific aims of this study are to make several protocol improvements based on our experiences from pilot testing to improve retention and evaluate this new protocol in a home training study with chronic stroke subjects.